BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:LIDS Seminar | Aaron D. Ames\, Caltech
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260915T211935Z
UID:tag:localist.com\,2008:EventInstance_51028900645251
DTSTART:20251020T200000Z
DTEND:20251020T210000Z
DESCRIPTION:Title: Foundations for Safe Autonomy: Why Learning Needs Contro
 l\n\n \n\nAbstract: With the rise of humanoids and the rapid deployment of
  learning across autonomy stacks\, the central question is: how can we tru
 st robots to operate safely around us? Despite impressive performance gain
 s\, learning at scale introduces fragility—making safety the key blocker
  to reliable deployment. This talk outlines the foundations for safe auton
 omy\, coupling learning with the formal guarantees from control theory. Th
 e cornerstone of this approach is Control Barrier Functions (CBFs)\, which
  encode safety as forward set invariance.  This leads to safety filters: r
 eal‑time wrappers that take desired commands—even from black‑box\, l
 earning‑enabled components—and minimally modify them (only when needed
 ) to keep the system safe.  These filters naturally sit within layered aut
 onomy stacks\, yielding a coherent architecture for trustworthy robots. Fi
 nally\, I will close the loop with learning: utilizing Lyapunov‑ and bar
 rier‑based reward shaping and shielding to enforce stability and safety 
 during training\, not just at deployment.  This framework for safe autonom
 y is grounded in—and will be illustrated by—extensive experimental res
 ults across diverse robotic platforms: ground vehicles\, drones\, aircraft
 \, legged and humanoid robots. \n\n \n\nBio: Aaron D. Ames is the Bren Pro
 fessor of Mechanical and Civil Engineering\, Control and Dynamical Systems
 \, and Aerospace at Caltech\, and the Director and Booth‑Kresa Leadershi
 p Chair of the Center for Autonomous Systems and Technologies (CAST). His 
 research centers on nonlinear control and its application to robotic syste
 ms—both formally and through experimental validation—with a special fo
 cus on legged and humanoid robots.  He pioneered Control Barrier Functions
  (CBFs) and safety filters for the safety‑critical control of highly dyn
 amic robots. An IEEE Fellow\, his recognitions include the NSF CAREER Awar
 d (2010)\, the Donald P. Eckman Award (2015)\, the Antonio Ruberti Young R
 esearcher Prize (2019)\, and more than twenty best‑paper awards\, includ
 ing ICRA Best Paper (2020\, 2023). He earned B.S./B.A. degrees in Mechanic
 al Engineering and Mathematics from the University of St. Thomas (2001) an
 d an M.A. in Mathematics and a Ph.D. in EECS from the University of Califo
 rnia\, Berkeley (2006). He was a postdoc at Caltech and held faculty posit
 ions at Texas A&M and Georgia Tech\, before joining Caltech in 2017.
GEO:42.361613;-71.092293
LOCATION:Building 45 (MIT Stephen A. Schwarzman College of Computing)\, 230
SUMMARY:LIDS Seminar | Aaron D. Ames\, Caltech
URL;VALUE=URI:https://calendar.mit.edu/event/lids-seminar-aaron-d-ames-calt
 ech
CATEGORIES:Conferences/Seminars/Lectures
END:VEVENT
END:VCALENDAR